DESCRIPTION: An avalanche of sequencing and genomic data has a potential to revolutionize our understanding of microbial biology and transform medical research. Accurate reconstructions of bacterial metabolism often provide the most direct route toward understanding the biology of sequenced species, their growth and environmental properties, and possible interactions with other species in metagenomic communities. Reconstructed metabolic networks can also guide the development of new antibacterial therapeutics. In this application we propose to use the GLOBUS framework to build an integrated, accurate, and fully probabilistic system for the annotation of bacterial metabolism; we will integrate into the framework key data modalities that, according to our preliminary results, will significantly improve the method's coverage and accuracy. Specifically, we will a.) integrate protein structural information into GLOBUS, using analyses of enzyme active sites; b.) extend the GLOBUS framework to accommodate metabolomics data by joint sampling of metabolomics and protein annotations; and c.) integrate into the annotation framework phenotypic information, including growth on multiple nutrient sources measured by the widely used BiOLOG platform for high-throughput bacterial phenotyping. We will implement and make the developed methodology publically available through a transparent web-based portal. This will make it possible for other researchers to use the GLOBUS methodology to annotate any sequenced bacterial species of interest. The portal will be transparent, enabling users to identify the sources of the predictions We will also obtain relevant phenotypic information from our experimental collaborators and use GLOBUS to generate accurate probabilistic annotations for all major bacterial species (~50 bacteria) that are pathogenic to humans.